Automatic call director (ACD) technology is a well known technology that: i) accepts incoming calls from calling clients; ii) queues calling clients on hold for connection to a service representative; and iii) when a service representative is available, connects the customer from the queue to the available service representative.
ACD systems also include expected wait time systems which, as a courtesy, notify a caller of the predicted time which the caller will need to wait in queue until a service representative is available.
Several known models exist for calculating estimated wait time. For example, in U.S. Pat. No. 5,506,898 to Costantini et al., a system for calculating an estimated wait time includes: i) calculating a rate of advance for each caller in the queue; ii) calculating a weighted average rate of advance; and calculating an estimated wait time for a caller as a function of the weighted average rate of advance and the caller's position in the queue.
In a variation, it is also known to calculate the predicted wait time as a function of the number of calls waiting in queue and a time factor—which is a value representing an average time for servicing clients.
Further yet, it is known to track the status of a service being provided to a caller who has connected to service representative to estimate when that particular call will be completed and the representative will be available to take the next call in the queue. For example, in U.S. Pat. No. 6,694,009 to Anderson et al., information is collected about a plurality of “points in the processing of a call”. This collected information is combined with historical information related to processing of calls to estimate how long it is likely to take until the call is completed.
It is also known to track a historical average time spent in queue by calls that have been processed by the queue and calculate predicted wait time as a function of historical wait time and the client's position in the queue.
A problem with existing estimated wait time calculation schemes is that they are all subject to significant inaccuracies, particularly if used in conjunction with a service representative work force that that changes over various time periods throughout the day.
For example, if the number of service representatives is increased, the number of customers that can be simultaneously serviced is also increased. This would likely result in a calculated expected wait time value which represents a time duration that is inaccurately longer than that actually experienced by the caller. The same is true in reverse. If the number of service representatives is decreased, the number of customers that can be simultaneously serviced is also decreased. This would likely result in a calculated expected wait time value which represents a time duration that is inaccurately shorter than that actually experienced by the caller.
What is needed is a system and method for improving the accuracy of a wait time calculation and, more particularly, improving the accuracy of a wait time calculation in an environment wherein a work force changes over various time periods through out the day. Further, what is needed is such a system that is useful for providing callers with a predicted wait time that reflects predicted changes in the availability of service representatives.